article Ryan Cooper
We are entering the holiday season, which typically means fewer races, more (over)eating, and my personal favorite— race season goal setting.
The off-season is also a time for reflection on the last season or seasons of effort. At Best Bike Split (BBS) we’re doing some of our own reflections. November marks three years since we became part of the TrainingPeaks family. Our little startup race day planning and prediction application has come such a long way and grown in ways and directions we had not originally conceived.
As part of that journey I think it’s important to explain how BBS came to be, some of the ways it can be used today for better season planning, and what our plans are for its future.
Four years ago I was a PhD student toiling away on somewhat uninteresting projects when I happened to stumble into the Cycling Aerodynamics and Performance Modeling Symposium/Webinar hosted by Jim Martin.
The presentations opened my eyes to the possibilities of applying my specific background of study to cycling. Interestingly enough, this past year I met Martin in Germany before the start of the 2017 Tour de France and was able to personally thank him for unknowingly creating the spark that led to BBS.
My first son was born on July 3, 2013, and during that next week (before the first individual time trial of the 2013 Tour de France) I worked—and helped with the baby—pretty nonstop to develop a prediction model.
Using prior years’ available data, courses, weather conditions, rider performance and any public power files, I created a physics-based model that included physiological constraints to estimate an athlete’s performance.
The night before the Tour stage I went to my wife with a list of predicted times from some of the expected top performers. The excitement on her face seven days post-baby was muted to say the least, but when times started coming in things started to change as the differences were just seconds off the model!
While this is interesting, it’s a ton of work to manually model things, and the outputs are definitely not sexy or really that useful. It wasn’t until I saw the stage winner Tony Martin’s race day plan for that specific race with every section broken down with targets and instructions that I knew what an app needed to do.
For me it’s not as much about the time prediction as much as it is about the plan: How do I race to my utmost ability on the specific course and day? With the growing prevalence of power meters it was obvious that a power plan was really the key. Knowing my programmatic and design limitations, I pitched the idea over fajitas to my longtime buddy and creative/dev extraordinaire Rich Harpel to make it a reality. A few months later he had made the first beta release of Best Bike Split.
Its funny how reaching a goal often takes a different path than originally envisioned. We were approached by a couple companies in our first six months of existence, but one meeting with Gear Fisher, CEO of Peaksware, sealed the deal of our vision and where we thought things could flourish.
The team at TrainingPeaks is outstanding, and a meeting with Chief Marketing Officer Jeremy Duerksen sparked what is probably the most important feature of BBS. He said the application is great if you know where you are racing and what your goal is—but what about those who don’t? What if we could let users turn the dials and see the impact? The light bulb went off and we started work on what has become the Time Analysis Tool.
How BBS Can Help You Set the Right Goal
When I used to race I would hear goals thrown out all the time:
“This year I’m going to do a five- hour IRONMAN bike split.”
“If I can just lose five pounds I know my times will be so much faster.”
“I should get a new bike or new set of wheels to improve my times.”
The problem with all of these is that there is usually an unknown basis to them. Let’s say you are racing IRONMAN Florida versus IRONMAN Wisconsin; the effort level to achieve a five-hour bike split is completely different on these courses.
These type of situations are where the Time Analysis Tool is invaluable. Now, not only can you show the expected time of these different courses, but you can vary power, weight and aerodynamics to see what is needed physiologically to achieve these goals.
Using this, you can then set season FTP or weight goals, or even decide which races might benefit your strengths and hide potentially weak areas. It’s the ultimate “what-if” planning analysis.
For even more fine-tuned goals, like many of our users do with their coach, a deeper study of the Time Analysis Tool can provide insight into the best places to attack a course in a road race, where a potential bike exchange in a hilly time trial might make sense, or even goal power for various sections of a long MTB race like the Leadville 100.
While these micro-goals and race strategy concepts do take more planning and analysis, the tools we provide allow anyone from a first time Gran Fondo rider, to the Tour de France teams, and even the coaches of the Olympic time trial champions to fine-tune their strategies for the specific goals of the event.
No matter the distance, every race is different. One IRONMAN can rarely be directly compared to another. The total elevation change and how that elevation is dispersed (i.e. frequency, length and grade of climbs), the wind and weather—even the number of technical turns all play a big role in your specific race.
When you don’t live and train in the conditions expected on race day, the challenge of fine tuning and conditioning your body for the challenges of the course and environment become very difficult. With BBS we have teamed with our partners at Wahoo, Garmin, TrainerRoad and Zwift to mimic the rigors of racing and training for the specific course and race day conditions on the trainer.
By loading a BBS race plan (or specific section of a race) into your indoor training platform of choice you can condition your legs to feel the power needed to race in the different wind conditions or climbing situations you will experience on race day for the estimated time it will take.
Scottish triathlete Graeme Stewart used this technique to prepare for the difficult sections of the 2015 Norseman Xtreme Triathlon. Jeff Agar was able to simulate the impact of pulling his son Johnny and a 60-pound Chariot while indoors in the cold winter months in Minnesota so that he would be prepared to do it during both IRONMAN and IRONMAN 70.3 races.
Because a race is typically not a perfectly steady-state effort, training specifically for the course on longer rides allows you to know exactly what efforts to expect and, more importantly, when in the race to expect them. Sprinkling these simulation efforts into your training provides not only a good race-pace workout, but also confidence that you will be ready for the day.
In the math-modeling world, the data going into the model must be of high quality to get optimal results. The concept of bad data in will always equal bad data out is truly apparent when you run a model and see results that are way out of scope from expectations (or previous experiences).
To this point we know that with so many variables going into Best Bike Split there are a lot of areas where inputs can be hard to fine tune. We have spent a lot of time refining our models behind the scenes to account for elevation data and weather data, but the biggest source of error for most models comes down to aerodynamic drag settings.
Without having access to a wind tunnel or velodrome for testing, coming up with a accurate drag number is a difficult task. To simplify this and allow for a more accurate model we have developed two methods for helping athletes and coaches dial in an athlete’s drag details.
The first is through modeling a previous race in BBS and comparing the model to the actual ride, in a section where you know the athlete was maintaining ideal aero form, using our Time Analysis Tool.
By setting the average power for that section of the model to that of the real ride and varying the drag slider until the times match, an athlete or coach can narrow down a better drag estimate for that position.
Our secondary method is a Beta tool we call Aero Analyzer, which analyzes the athlete’s actual ride file to determine an estimate of both an aero and relaxed position on the bike. This tool is best used with a recent race result or a test ride of 30 minutes or longer where speed and power were varied throughout the ride. As we continue to iterate the tool we expect it will allow for even more analysis and data enhancement.
Best Bike Split has come a very long way since our humble beginnings, but as a new season approaches we are just about to explore our 2018 goals. This year is about helping athletes and coaches achieve the most accurate model possible and highlighting areas of strengths and potential weaknesses.
To do this we are working toward completing our original BBS vision, which not only includes race prediction, training specifically for the course and conditions, and providing a optimal race day power plan, but ultimately completes the feedback loop post-race to enhance future performances as well.
The goals we set today will help us achieve our dream of tomorrow—which is as true for us as it is for all you coaches and athletes out there.
Ryan Cooper is the Chief Scientist at TrainingPeaks and Co-founder of Best Bike Split. He has worked and consulted with multiple World, Olympic, and IRONMAN champions, as well as teams including UnitedHealth Care, Dimension Data, Cannondale Drapac, Orica Scott, BMC, Trek, and Sky. His main mission is spreading the metrics-based training approach of TrainingPeaks and the predictive race day analytics provided by Best Bike Split. Learn more at TrainingPeaks.com and BestBikeSplit.com.
You must be logged in to post a comment.
